EP2998981A1 - Graphoepitaxy method for creating patterns on the surface of a substrate - Google Patents

Abstract

The invention relates to a process for producing patterns on the surface of a substrate (2) by self-assembly of block copolymer, comprising the following steps:
forming an assembly guide (1) on first and second zones (20a, 20b) of the substrate, the assembly guide having, with respect to a reference surface (1a), openings with a degree of opening in the first zone (20a) greater than that of the second zone (20b);
depositing a layer of block copolymer (4) on the substrate (2), so that the layer of block copolymer (4) completely fills the assembly guide (1) and forms an extra thickness on the reference surface ( 1 a);
- assembling (F4) the block copolymer, from which there results an organized portion (4a) of the block copolymer layer within the openings of the assembly guide (1);
uniformly thinning the block copolymer layer (4) to a thickness corresponding to the organized portion (4a) of the block copolymer layer;
removing one of the phases of the assembled block copolymer, resulting in a plurality of initial patterns extending into the block copolymer layer; and
- Transfer the initial patterns of the block copolymer layer into the substrate (2) to form said final patterns.

Description

TECHNICAL AREA

The present invention relates to directed self-assembly techniques of Directed Self-Assembly (DSA) to generate very high resolution and density patterns. More particularly, the invention relates to a method of making patterns on the surface of a substrate, such as contact holes or trenches, using block copolymers.

STATE OF THE ART

Process requirements for making nano-sized patterns have increased significantly in recent years, due to the trend toward miniaturization of electronic components.

Initially, the patterns were made by optical projection lithography. In this method of realization, a photosensitive layer is deposited on a substrate, then exposed to an ultraviolet beam through a mask defining the patterns. The size of the smallest achievable pattern (also called critical dimension "CD" for "Critical Dimension") is closely related to the wavelength of the beam used: the shorter the wavelength, the more the pattern can be made. be finer and more integration density of these patterns in an integrated circuit can be important.

The ultraviolet beams used in photolithography conventionally have a wavelength of 193 nm or 248 nm. This method of defining patterns has the advantage of being well controlled and less expensive than other methods, including electronic lithography processes. However, with such wavelengths, the resolution of the exposure tool is limited.

In order to achieve finer patterns, better solved, and thus to further increase the integration density, new processes have been developed, such as double-patterned photolithography ("double-patterning"), Extreme UltraViolet lithography (EUV). and electron beam lithography (e-beam).

Emerging lithography technologies also include Directed Self-Assembly (DSA) directed self-assembly techniques. Block copolymers are polymers in which two repeating units, a monomer A and a monomer B, form chains linked together by a covalent bond. When the chains are given sufficient mobility, for example by heating these block copolymers, the chain A and the chain B tend to separate into phases or blocks and to reorganize under specific conformations, which depend in particular on the ratio between the monomer A and the monomer B. Depending on this ratio, it is possible, for example, to have spheres of the monomer A in a matrix of the monomer B, or else rolls of A in a matrix of B, or else lamellae of A and strips of B interposed. The size of the domains of the block A (respectively of the block B) is directly proportional to the length of the chains of the monomer A (respectively of the monomer B). Block copolymers thus have the property of forming patterns that can be controlled by the ratio of monomers.

The known techniques for self-assembly of block copolymers can be grouped into two categories, grapho-epitaxy and chemi-epitaxy, both described in detail in the article [" Guided self-assembly of block-copolymer for CMOS technology: a comparative study between grapho-epitaxy and surface chemical modification, L. Oria et al., SPIE 2011, Vol 7970-24 ].

Chemi-epitaxy consists of modifying the chemical properties of certain portions of the substrate, to force the organization of the copolymer blocks between these portions. These portions of the chemically modified substrate are typically delimited by a photolithography step followed by a plasma step.

Alternatively, grapho-epitaxy consists of forming primary patterns called guides on the surface of a substrate, these patterns delimiting zones within which a layer of block copolymer is deposited. The guide patterns control the organization of the copolymer blocks to form higher resolution secondary patterns within these areas. The guide patterns are conventionally formed by photolithography in a resin layer, and possibly transferred into a hard mask.

The grapho-epitaxy technique has recently been used to form contact holes in an integrated circuit. After deposition and assembly of the block copolymer in the guiding units, the secondary units are developed by selectively removing one of the two blocks of the copolymer (for example the rolls of A), thereby forming holes in the remaining layer of copolymer ( the matrix of B). Then, these holes are etched onto the surface of the substrate, usually in a dielectric layer.

With this technique, we can reduce the size of the patterns (application called "contact shrink" in English), the secondary pattern is smaller than the primary pattern, i.e. the guide pattern. There is then only one contact hole per guide pattern. It is also possible to multiply the number of patterns by forming several contact holes per guide pattern ("multiplication contact").

As discussed in article [" Etch challenges for DSA implementation in CMOS via patterning ", P. Pimenta Barros et al., SPIE Proceedings Vol 9054, March 2014 ], the current methods of grapho-epitaxy are dependent on the density of the guide patterns on the substrate. Indeed, since the filling of the guiding motifs is done by spin coating, the thickness of the block copolymer layer inside a guiding pattern depends on its area, its aspect ratio (depth / width) and the number of guide patterns nearby. Therefore, for guiding patterns of the same dimensions, the thickness of the copolymer layer in an insulated pattern is greater than the thickness obtained in a plurality of closely spaced patterns.

However, the thickness of the copolymer layer in the assembly guides affects the transfer of the patterns by etching, because the copolymer layer serves as an etching mask. If in some assembly guides the copolymer thickness is too small, an increase in the critical dimension of the patterns can occur during their transfer, because the etching mask is insufficient. Conversely, when the copolymer thickness is too great, contact holes may be missing, their transfer having failed.

The thickness of the copolymer layer also affects the step of self-assembly of the block copolymer. In fact, a too small or too large thickness of the copolymer layer in the guides can cause a bad organization of the blocks. In particular, some polymer patterns may not orient perpendicularly to the substrate. In general, these assembly defects relate to insulated guides, where the thickness of block copolymer is the largest.

Thus, since the thickness of the layer of block copolymer varies within the guiding units of the same substrate (for example according to their density), it is rare to obtain the assembly and transfer of all the secondary units. with the same performance, especially in terms of critical dimension.

SUMMARY OF THE INVENTION

There is therefore a need to provide a method for producing final patterns on the surface of a substrate by self-assembly of block copolymer that is independent of the configuration of the assembly guide, that is to say a process which allows to obtain patterns with the same performance regardless of the distribution and dimensions of the guide patterns.

• former un guide d'assemblage sur des première et seconde zones du substrat, le guide d'assemblage présentant par rapport à une surface de référence des ouvertures avec un taux d'ouverture dans la première zone supérieur à celui de la seconde zone ;
• déposer une couche de copolymère à blocs sur le substrat, de sorte que la couche de copolymère à blocs remplisse entièrement le guide d'assemblage et forme une surépaisseur sur la surface de référence ;
• assembler le copolymère à blocs, d'où il résulte une portion organisée de la couche de copolymère à blocs à l'intérieur des ouvertures du guide d'assemblage ;
• amincir uniformément la couche de copolymère à blocs, jusqu'à atteindre une épaisseur correspondant à la portion organisée de la couche de copolymère ;
• éliminer une des phases du copolymère à blocs assemblé, d'où il résulte une pluralité de motifs initiaux s'étendant dans la couche de copolymère à blocs ; et
• transférer les motifs initiaux de la couche de copolymère à blocs dans le substrat pour former lesdits motifs finaux.

According to the invention, there is a tendency to satisfy this need by providing the following steps:

• forming an assembly guide on first and second zones of the substrate, the assembly guide having openings with an opening ratio in the first zone greater than that of the second zone relative to a reference surface;
• depositing a layer of block copolymer on the substrate, so that the block copolymer layer completely fills the assembly guide and forms an extra thickness on the reference surface;
• assembling the block copolymer resulting in an organized portion of the block copolymer layer within the openings of the assembly guide;
• uniformly thinning the block copolymer layer to a thickness corresponding to the organized portion of the copolymer layer;
• removing one of the phases of the assembled block copolymer, resulting in a plurality of initial patterns extending into the block copolymer layer; and
• transferring the initial patterns of the block copolymer layer into the substrate to form said final patterns.

The assembly guide may designate a layer (or a stack of several layers) in which openings or cavities are formed, for example by photolithography. According to an alternative embodiment, the guide may be made in the form of openings directly in the substrate. These openings form guide patterns in which the block copolymer is deposited.

The ratio of the area of the openings in the guide, and the total area of a field containing these openings, for example 2 μm × 2 μm, is referred to above as "opening rate of the assembly guide". When all the openings of the substrate guide patterns have the same area, an aperture rate in a first area of the substrate greater than that in a second area of the substrate equals a greater opening density in the first area than in the second zone. It is also possible to achieve a higher opening rate in the first zone, without necessarily having a greater number of openings, when the openings of the first zone have larger areas than those of the second zone.

The deposition step, hereinafter referred to as a "planarization step", makes it possible to obtain a layer of block copolymer, which protrudes from the assembly guide (forming a excess thickness above the upper face of the guide, said reference surface) and whose surface is substantially flat.

After the step of assembling the block copolymer, a portion of the block copolymer layer located inside the openings of the assembly guide is found organized according to the desired pattern. In general, it is sought to obtain patterns oriented perpendicular to the plane of the substrate and, advantageously, arranged periodically. The remaining portion of the block copolymer layer (i.e. the excess thickness) may be oriented or not (i.e. disorganized). If it is oriented, it can be oriented in the same direction as the patterns of the organized portion or another orientation. The remaining portion generally has defects. By default, we mean any break in the periodic network. These defects are erased by uniformly thinning the block copolymer layer, to keep only the properly organized copolymer at the bottom of the guide.

Thus, thanks to the planarization step, followed by the assembly step and the thinning step, an organized copolymer layer of the same thickness is obtained in all the substrate guide patterns, whatever their size and their distribution in the first and second zones. We can then remove one of the blocks of the organized copolymer and transfer the patterns generated by this withdrawal in the substrate in the same way for all the guide patterns, unlike the processes of the prior art where there is no stage planarization, or thinning step.

The portion of the layer that is not located inside the openings of the assembly guide and organized according to the desired pattern - in which the copolymer domains are arbitrary, for example oriented parallel to the substrate (and not perpendicularly) -, does not affect the transfer, since it is first removed during the thinning stage.

The thinning step of the block copolymer layer, carried out for example by plasma etching and / or chemical-mechanical polishing, advantageously takes place after the copolymer assembly step. Thus, the monomer blocks are separated and frozen during the step of thinning the layer. There is therefore no risk of disturbing the organization of the domains of the copolymer. The thinning step may nevertheless be performed before the assembly step or in part before and partly after.

The thinning step and the step of removing one of the phases of the assembled block copolymer are advantageously carried out in the same plasma etching frame, or simultaneously using a single plasma chosen from Ar / O ₂ , C _x F _y , SF ₆ , N ₂ / H ₂ , CO / O ₂ , CO / H ₂ , CH ₄ / O ₂ , C _x F _y / O ₂ , CH _x F _y / O ₂ , C _x F _y / H ₂ , CH _x F _y / H ₂ and C _x H _y / H ₂ , either by a succession of steps using plasmas of different natures or by alternation of at least two plasmas.

Preferably, the deposition of the block copolymer layer is carried out by spin coating.

• le guide d'assemblage est formé d'un masque disposé sur le substrat et texturé par photolithographie et/ou gravure ou réalisé directement dans le substrat par photolithographie et/ou gravure ;
• ce masque comprend une première couche carbonée en contact avec le substrat et une seconde couche anti-réflective riche en silicium disposée sur la première couche carbonée, ladite couche anti-réflective riche en silicium servant de couche d'arrêt lors de l'amincissement de la couche de copolymère à blocs assemblé ;
• le guide d'assemblage comporte une pluralité d'ouvertures formant des motifs de guidage répartis entre les première et seconde zones du substrat ;
• les motifs de guidage ont une hauteur comprise entre 5 nm et 1000 nm ;
• la surface des motifs de guidage présente une affinité particulière pour un ou plusieurs des blocs du copolymère à blocs ;
• le fond des motifs de guidage est neutre par rapport aux blocs du copolymère à blocs tandis que les flancs des motifs de guidage présentent une affinité particulière pour un ou plusieurs des blocs du copolymère à blocs ;
• la surface des motifs de guidage est neutre par rapport aux blocs du copolymère à blocs ;
• l'affinité particulière du guide d'assemblage pour un ou plusieurs blocs du copolymère à blocs est obtenue par le greffage d'un ou plusieurs homopolymères ; et
• la neutralisation du guide d'assemblage est obtenue par le greffage d'un copolymère statistique.

The method according to the invention may also have one or more of the following characteristics, considered individually or in any technically possible combination:

• the assembly guide is formed of a mask disposed on the substrate and textured by photolithography and / or etching or made directly in the substrate by photolithography and / or etching;
• this mask comprises a first carbon layer in contact with the substrate and a second silicon-rich anti-reflective layer disposed on the first carbon layer, said silicon-rich anti-reflective layer serving as a stop layer during the thinning of the block copolymer layer assembled;
• the assembly guide comprises a plurality of openings forming guide patterns distributed between the first and second zones of the substrate;
• the guiding units have a height of between 5 nm and 1000 nm;
• the surface of the guide patterns has a particular affinity for one or more blocks of the block copolymer;
• the bottom of the guide patterns is neutral with respect to the blocks of the copolymer while the flanks of the guiding units have a particular affinity for one or more blocks of the block copolymer;
• the surface of the guide patterns is neutral with respect to the blocks of the block copolymer;
• the particular affinity of the assembly guide for one or more blocks of the block copolymer is obtained by grafting one or more homopolymers; and
• the neutralization of the assembly guide is obtained by grafting a random copolymer.

BRIEF DESCRIPTION OF THE FIGURES

• les figures 1A à 1G représentent par des vues de coupe des étapes d'un procédé pour réaliser des motifs à la surface d'un substrat, selon un mode de mise en oeuvre préférentiel de l'invention ;
• la figure 2 représente la variation d'épaisseur de copolymère à blocs au sein d'un guide d'assemblage en fonction de la densité des motifs dans le guide, après l'étape de dépôt selon un procédé de l'art antérieur (essai n°1) et après l'étape d'amincissement du procédé selon l'invention (essais n°2 et 3) ; et
• les figures 3A et 3B représentent schématiquement une couche de copolymère à blocs au sein d'un guide d'assemblage, respectivement après l'étape de dépôt selon un procédé de l'art antérieur et après l'étape d'amincissement du procédé selon l'invention.

Other features and advantages of the invention will emerge clearly from the description which is given below, as an indication and in no way limiting, with reference to the appended figures, among which:

• the Figures 1A to 1G represent, by cutting views, the steps of a method for producing patterns on the surface of a substrate, according to a preferred embodiment of the invention;
• the figure 2 represents the variation of block copolymer thickness within an assembly guide as a function of the density of the patterns in the guide, after the deposition step according to a method of the prior art (test No. 1) and after the thinning step of the process according to the invention (tests Nos. 2 and 3); and
• the Figures 3A and 3B schematically represent a block copolymer layer within an assembly guide, respectively after the deposition step according to a method of the prior art and after the thinning step of the method according to the invention.

For the sake of clarity, identical or similar elements are identified by identical reference signs throughout the figures.

DETAILED DESCRIPTION OF AT LEAST ONE EMBODIMENT

A preferred mode of implementation of the method according to the invention will now be described with reference to Figures 1A to 1G .

The method firstly comprises a step F1 represented in FIG. Figure 1A during which an assembly guide 1 is formed on the surface of a substrate 2. The assembly guide 1 comprises openings 10 which extend from the upper face 1a of the guide to the substrate 2. These openings 10 correspond to to the guide patterns in which the block copolymer will be deposited.

The substrate 2 comprises at least two zones 20a and 20b on which the guide patterns 10 are distributed. The distribution of the patterns 10 within the assembly guide 1 is such that the guide has an opening rate in the upper zone 20a. to that in zone 20b. In the example of the Figure 1A this is schematized by three openings 10 in the zone 20a, while the zone 20b has only one, all the openings 10 having substantially the same dimensions. Here, one wishes to form three times more patterns in the zone 20a than in the zone 20b.

One could also consider a zone 20a comprising as many openings 10 as the zone 20b, but openings whose dimensions are greater than those of the zone 20b. This case occurs in particular when it is desired to form, on the same substrate, several patterns (for example contact holes) by opening in zone 20a ("multiplication contact") and a single pattern by opening in zone 20b ( "Contact shrinking").

The assembly guide 1 may be formed of a hard mask which covers the substrate and consists of one or more layers. The guide patterns 10 in the mask are then obtained by etching these different layers. Typically, a layer of resin is deposited on the hard mask, then printed by means of a photolithography (at a wavelength of 193 nm for example), an UltraViolet Extreme Lithography (EUV), a lithography by electron beam ("e-beam") or any other lithographic technique (nano-printing, multiple exposure photolithography ...). Then, the guide patterns printed in the resin are transferred by etching into the layers of the hard mask.

Preferably, the hard mask is a bilayer stack comprising a first carbon layer 11 ("Spin On Carbon", SOC), in contact with the substrate 2, and a second anti-reflective layer 12 disposed on the carbon layer 11. The layer anti-reflective 12 is, for example, a silicon-rich coating ("Silicon Anti-Reflective Coating", SiARC). This bilayer stack notably makes it possible to adapt the optical refractive index of the substrate, avoiding the creation of standing waves during the photolithography step, and has good selectivity in the face of etching of the block copolymer.

In an alternative embodiment (not shown), the assembly guide may be formed of a resin layer disposed directly on the substrate, the guide patterns having been printed in the resin by means of any of the aforementioned lithography, for example by photolithography (insolation and development of a portion of the resin). This resin may in particular be a negative tone development resist. A solvent is then used to remove the unexposed regions of the resin, while the exposed regions remain on the substrate.

Seen from above, the guide patterns 10 of the guide 1 can have different shapes, in particular rectangular, circular or elliptical, and spacings or no repetition ("pitch" in English) variables, in order to obtain a reduction and / or a multiplication of the contact holes ("contact shrink" and / or "contact multiplication"). They can also form networks of lines. The thickness of the guide 1, that is to say the height of the patterns 10, is advantageously between 5 nm and 1000 nm, typically between 20 nm and 300 nm.

In this preferred embodiment, a surface preparation step may be performed to promote the disposition of the domains of the block copolymer, for example perpendicular to the surface on which it will be deposited. This optional step F2 is represented at Figure 1B .

• l'ensemble de la surface des cavités 10 présente une affinité particulière pour l'un ou plusieurs des blocs du copolymère ;
• les flancs des cavités 10 présentent une affinité particulière pour un ou plusieurs des blocs du copolymère et le fond des cavités 10 est neutre par rapport aux blocs du copolymère ;
• l'ensemble de la surface des cavités 10 est neutre par rapport aux blocs du copolymère.

Three modes of preparation can be envisaged to obtain three types of configuration:

• the entire surface of the cavities has a particular affinity for one or more of the blocks of the copolymer;
• the flanks of the cavities 10 have a particular affinity for one or more of the blocks of the copolymer and the bottom of the cavities 10 is neutral with respect to the blocks of the copolymer;
• the entire surface of the cavities 10 is neutral with respect to the blocks of the copolymer.

In order for a surface to have a particular affinity for one or more copolymer blocks, one or more suitable homopolymers can be grafted onto it. Surface neutralization can be obtained by grafting a random copolymer. As is visible on the figure 2 a layer 3 made of homopolymer or random copolymer is then grafted onto the bottom and / or the side walls of the guiding units 10. The grafted layer 3 typically has a thickness of 2 nm to 20 nm. The grafting of the layer 3 is conventionally carried out in three steps: deposition of the material (homopolymer or random copolymer), annealing and rinsing.

Other ways of neutralizing the substrate can be envisaged, in particular by plasma treatment, UV treatment, ion implantation or deposition of a self-assembled monolayer (SAM). The substrate may also be chosen to promote the arrangement of the copolymer domains. It is therefore not mandatory to use a neutralization layer.

At step F3 of the figure 1C a layer of block copolymer 4 is deposited in the assembly guide 1. The difference in the opening rate of the assembly guide in several zones of the substrate causes, in the method of the prior art, a difference in thickness of the copolymer layer inside the guide. Conversely, in the process of Figures 1A to 1F the deposition conditions are chosen so that the layer of block copolymer 4 has, in each zone 20a, 20b of the substrate, substantially the same thickness. This requires that the layer 4 protrudes from the assembly guide 1. Thus, the guide patterns 10 are not only completely filled, but in addition, the layer of block copolymer 4 is in excess thickness above of the guide 1.

However, in conventional DSA self-assembly processes, it is avoided to deposit a layer of too great thickness because, as will be described later, the assembly of a block copolymer outside a guide causes defects orientation.

The block copolymer of layer 4 can be a di-block copolymer (two monomers A and B) or multi-block copolymer (more than two monomers), a mixture of polymers, a mixture of copolymers or the mixture of a copolymer and a homopolymer. It can be of any morphology, for example spherical, cylindrical, gyroid, lamellar, depending on the proportion between the monomer blocks.

The material of layer 4 is, for example, a diblock copolymer containing a styrene derivative (typically polystyrene, PS) and a methalcrylate derivative (typically polymethyl methacrylate, PMMA). With 30% by weight of PMMA and 70% by weight of PS, the PMMA domains are, after assembly, in the form of cylinders contained in a PS matrix. Of course, other proportions between the monomers can be envisaged, in order to obtain domains of different geometry (lamellar, spherical ...).

The deposition of the block copolymer layer is preferably carried out by spin coating (or "spin coating" according to the English term). A solution containing a solvent (for example toluene or propylene glycol monomethyl ether (PGMEA)) and the copolymer material (for example PS-b-PMMA) is spread on the substrate by centrifugal force. This deposition mode is particularly suitable for flattening a surface topography, such as that generated by the cavities of the assembly guide 1 on the substrate 2.

où γ est la tension de surface, h est la hauteur de la cavité, ρ la densité du polymère, ω la vitesse de rotation du substrat et r la position radiale de la cavité sur le substrat.The planarization of a surface topography by spin coating has been the subject of numerous studies. In particular, in the article [" Planarization of Substrate Topography by Spin Coating, J. Electrochem Soc 1987 Volume 134, issue 8, 2030-2037 ], The Stillwagon et al. have shown that the efficiency of this planarization depends on the dimensions of the cavity to be filled (cf. Fig.1C , width w and height h), the thickness (h _f ) of the deposited polymer film and the viscosity of the polymer. According to LE Stillwagon et al., The surface becomes flat when the capillary forces are equal to the centrifugal force during operation of the spin. It can thus be defined a critical cavity width w _c below which the polymer surface is flat immediately after spin coating. This critical width is written: $$w_{\mathrm{vs}}={\left(\frac{\mathrm{16}\mathrm{\gamma h}}{{\mathrm{\rho \omega }}^{\mathrm{2}}r}\right)}^{\frac{1}{3}}$$

where γ is the surface tension, h is the height of the cavity, ρ the density of the polymer, ω the rotational speed of the substrate and r the radial position of the cavity on the substrate.

Thus, all cavities having a width less than the critical width w _c can be smoothed by the spin coating. It can be concluded that a narrow cavity (low w) is easier to flatten than a wide cavity (large w). On the other hand, a dense pattern field can be considered during spin coating as a single, wide cavity. In the assembly guide example of the figure 1C the three patterns of the zone 20a are sufficiently close together to be seen as a single cavity of dimensions three times larger than the pattern of the zone 20b. It is therefore easier to flatten isolated contacts, such as the pattern of zone 20b, than a dense pattern field (zone 20a).

où ω est la vitesse de rotation du substrat, η la viscosité du polymère, γ la tension de surface et h_f l'épaisseur du film de polymère déposé sur le substrat.Equation (1) above does not take into account the viscosity of the polymeric material because it is based on the assumption that the spin-on material is a Newtonian fluid, i.e. where the viscosity does not vary with the rate of deformation or with time. According to LE Stillwagon et al., A second equation can then be introduced to take into account the fact that the response of a polymeric material subjected to a force can spread over time. This equation defines the time t _p necessary to completely fill the cavity with a viscous polymer material after stopping the spin and is written: $$t_p=\frac{3}{32}\frac{{\mathrm{<figure-callout\; id="4"\; label="\omega "\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\omega </figure-callout>}}^{\mathrm{4}}\mathrm{<figure-callout\; id="3"\; label="\eta \; \gamma h\; f"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\eta \; </figure-callout>}}{{\mathrm{\gamma h}}_f^{}}$$

where ω is the rotation speed of the substrate, η the viscosity of the polymer, γ the tension of surface and h _f the thickness of the polymer film deposited on the substrate.

It is found that the higher the viscous material material (high η), the longer the waiting time t _p is. In addition, the waiting time t _p increases when the thickness of the deposited layer decreases. It is therefore difficult to obtain a thin flat layer with a highly viscous material by spin coating.

Thanks to the teachings of this article, and taking into account the application envisaged here, it is possible to form a layer of block copolymer 4 of substantially identical thickness on all the zones of the substrate, including the zones 20a and 20b having different opening rates. The viscosity of the solution containing the block copolymer is indeed low (close to that of the dissolution solvent), for example less than 1.3 cP. In addition, the thicknesses of the layer 4 deposited in step F3 are much greater than in the method of the prior art (because the guide 1 is exceeded), so that the flat layer 4 can be obtained almost instantaneously after stopping the spin.

The thickness of the copolymer layer 4 deposited on the substrate 2 is advantageously between 1 and 2 times the thickness of the assembly guide 1. The deposition conditions, such as the rotational speed, the weight percentage of the copolymer The amount of copolymer necessary to completely fill the guide depends, of course, on the number and size of the guide patterns formed on the substrate, in particular their height. Various morphological characterization techniques, such as the profilometer, make it possible to verify that the layer 4 is flat on the surface, which means that all the guide patterns 10 are filled. Several examples of these deposit conditions will be given later.

The planarization rate obtained as a function of the thickness of the deposited block copolymer has been studied, in particular in the article [ A Parametric Study of Spin Coating over Topography, Ind. Eng. Chem. Res., Vol. 37, No. 6, pp. 2223-2227, 1998 ]. For step F3 of the figure 1C , the flatness of the copolymer layer at blocks 4 is considered satisfactory when the planarization rate (as defined in this article) is greater than 50%, and advantageously greater than 75%.

• l'épaisseur d'un film en polymère en fonction de la vitesse de rotation, pour une viscosité donnée ;
• l'épaisseur du film en fonction de la concentration en polymère dans la solution de dépôt ; et
• la viscosité en fonction de la concentration en polymère dans la solution.

The conditions for obtaining and controlling the deposited thickness are determinable from experimental laws giving:

• the thickness of a polymer film as a function of the speed of rotation, for a given viscosity;
• the thickness of the film as a function of the polymer concentration in the deposition solution; and
• the viscosity as a function of the polymer concentration in the solution.

These laws can be taken from Article [ "Characteristics of resist films produced by spinning", J. Appl. Phys., Vol. 49, No.7, July 1978 ].

The figure 1D represents the F4 assembly step of the block copolymer, after the deposition of the layer 4 in step F3. After assembly, there are two portions of the copolymer layer: a portion 4a located at the bottom of the assembly guide 1 and a portion 4b located on the portion 4a and substantially corresponding to the extra thickness of the copolymer layer 4 above the assembly guide 1.

In portion 4b, the domains of the copolymer (here PS-b-PMMA) are generally misdirected. The domains of PMMA and PS can be randomly arranged. For example, surface defects of the mixed orientation type, that is to say a mixture of PMMA cylinders oriented parallel and perpendicular to the substrate 2, are observed.

On the other hand, surprisingly, in the portion 4a, the domains of the PMMA block and the PS block are correctly oriented. With the copolymer example mentioned above, the PMMA domains are in the form of cylinders 41 perpendicular to the substrate 2, each cylinder 41 being surrounded by a matrix 42 of polystyrene.

As is the case on the figure 1D , the upper face of the organized portion 4a may not correspond to the surface of the assembly guide 1 (reference surface 1a). For example, for an assembly guide 1 formed of a stack SiARC / SOC 125 nm thick, the organized portion 4a of the PS- b -PMMA copolymer layer will reach 70 nm thick.

The self-assembly of the blocks is preferably carried out by means of a heat treatment or annealing. The temperature and duration of the annealing are advantageously optimized as a function of the thickness of the copolymer layer 4. The thicker the layer 4, the greater the duration and / or the annealing temperature is important. For example, for a layer thickness of about 50 nm, an annealing of 2 min at 250 ° C. can be carried out in the case of a PS- b- PMMA copolymer with a period of about 35 nm. The assembly of the block copolymer may alternatively be obtained by an ultrasonic treatment or by a so-called solvent annealing operation, rather than by a heat treatment, or by a combination of these two techniques, or by any other technique known to those skilled in the art.

Step F5 of the figure 1E consists in etching the portion 4b of the copolymer layer located on the surface of the assembly guide 1, until reaching the organized portion 4a located inside the guide 1. This etching is performed directionally and uniformly on the plate such that the copolymer layer thus thinned has the same thickness at any point of the substrate 2, and therefore in the areas 20a and 20b where the distribution and / or the dimensions of the guide patterns 10 differ. For example, anisotropic etching can be performed.

In one embodiment of step F5, the thinning of the block copolymer layer 4 is performed by chemical mechanical planarization (CMP).

In an alternative embodiment, the thinning of the copolymer layer 4 is performed by plasma etching. The SiARC layer 12 can advantageously serve as a stop layer for etching. It then allows better control of the thinning and avoids the deterioration of the guide patterns 10. In the case of an integration where the guide does not allow to be used as a stop layer (example of the resin guide), it it is possible to create this stop layer before the deposition of the block copolymer, for example by densifying the resin constituting the guide or by depositing an oxide on the guide. The deposition of the oxide can be done by different deposition techniques (PECVD, ALD, PEALD ...) in order to be compatible with the material constituting the guide in terms of deposition temperature. For example, a layer of silicon dioxide SiO ₂ may be deposited on the resin guide by PEALD ("Plasma-Enhanced Atomic Layer Deposition") at a temperature of 50 ° C or the resin of the guide can be densified during 50 seconds by an HBr plasma (100 sccm) under a pressure of 5 mTorr and with a power of 1200 W. The oxide is an easily integrable material but it can be replaced by any other material fulfilling the role of stop layer during thinning of the polymer and whose deposition is compatible with the guide.

The gases that can be used to etch block copolymers are, for example, O ₂ , CO, CO ₂ , H ₂ , N ₂ .... These gases can be combined with one another (eg CO / O ₂ , CO / H ₂ ...) or with other inert gases (eg Ar / O ₂ , Ar / N ₂ , He / O ₂ ...), polymerizing (ie carbonated, eg CH ₄ / O ₂ ) or fluorinated (C _x F _y / O ₂ , CH _x F _y / O ₂ , C _x F _y / H ₂ , CH _x F _y / H ₂ ...). A mixture of several types of gas can also be used, as well as other gas chemistries: C _x F _y , SF ₆ ...

For example, the plasma is generated from a mixture of Ar / O ₂ gas, at a pressure of 10 mTorr, with a power of 220 W and a bias voltage of 100 V. The amount of oxygen represents between 9% and 100% of the mixture. It is thus possible to etch the copolymer PS- b -PMMA selectively with respect to SiARC (a selectivity greater than 20, for example of the order of 40, can thus be achieved).

Unlike chemical mechanical polishing, plasma etching has the advantage of etching the copolymer without damaging the assembly guide 1. Indeed, given that the organized portion 4a of the block copolymer is within the guide 1, the chemical mechanical polishing has the effect of planing the guide at the same time as the block copolymer to achieve the desired thickness, for example 70 nm in a guide of 125 nm in height. Conversely, a plasma etching process can leave the assembly guide intact, because of its large size. selectivity with respect to the material (s) which constitutes (s) the assembly guide.

Finally, in another variant embodiment, the thinning of the copolymer layer is carried out wet, with or without the presence of ultraviolet exposure. Liquid solutions for etching block copolymers are solvents, such as PGMEA, toluene and cyclohexane. The chosen solution does not aggravate the assembly guide. In order to better control the thinning and the thickness of block copolymer remaining in the guide, the surface of the copolymer layer may be previously modified by a plasma using a gas such as He, Ar, HBr, N ₂ , CH ₄ , CO, H ₂ , C _x F _y , C _x H _y , CH _x F _y ... or a combination of several of these gases (eg Ar / H ₂ , He / CH ₄ ). The portion thus modified by plasma is then selectively removed by wet leaving in the guide only the unmodified portion. For example, the chemical modification plasma is generated from Ar gas at a pressure of 10 mTorr, with a power of 500 W and a bias voltage of 300 V.

Still in order to better control the thinning and the thickness of block copolymer remaining in the guide, several plasma etching and wet etching steps can alternate in order to have an accurate etching of the copolymer layer. This approach takes up the principle of the new ALE (Atomic Layer Etching) etching techniques, which rely on plasma / deposition etching cycles or plasma / wet etching cycles, on the use of pulsed gas or on use of RF voltage during plasma etching. All of these techniques can be used during the thinning steps to have better control of the etched block copolymer thickness and the one remaining in the guide.

It is preferable, in the case of thinning by plasma etching in particular, to carry out the assembly of the block copolymer before the thinning of layer 4, and not after, since the thinning could alter one of the two phases of the copolymer, so that it becomes difficult to separate these two phases.

The method then comprises a step F6 (cf. fig.1F ) elimination of one of the phases of the assembled block copolymer layer, here PMMA cylinders 41. For this, one can in particular immerse the remaining portion 4a of the block copolymer layer in 99% concentrated acetic acid for 10 minutes. Prior UV exposure can also be achieved.

Patterns 43 are then obtained in a residual layer 42 composed solely of the other phase of the copolymer, polystyrene. The resolution of these secondary reasons 43 is very high because of the same order of magnitude as the PMMA domains.

Preferably, the removal of the PMMA cylinders (step F6) is performed in the same plasma etching frame as the thinning of the copolymer layer (step F5).

A first strategy consists of carrying out both steps simultaneously, using a single plasma with gas mixtures (reactive / inert gas or reactive / polymerizing gas) such as Ar / O ₂ , C _x F _y , SF ₆ , N ₂ / H ₂ , CO / O ₂ , CO / H ₂ , CH ₄ / O ₂ , C _x F _y / O ₂ , CH _x F _y / O ₂ , C _x F _y / H ₂ , CH _x F _y / H ₂ and C _x H _y / H ₂ . To be able to achieve with the same plasma these two steps, the selectivity between polystyrene and PMMA during this etching is advantageously between 1 and 5.

A second strategy is to perform several successive etching steps within the same equipment, first to thin the block copolymer layer, then remove the PMMA selectively relative to the PS, with etching chemistries different between the two steps. For example, thinning can be achieved with a CF ₄ plasma (no selectivity between PMMA and PS) and PMMA removal by means of a CO / O ₂ plasma (PMMA / PS> 10 selectivity). allows independent control of the thinning and removal of PMMA by plasma etching. It is also possible to proceed by alternating at least two plasmas to further improve this control, that is to say by performing successive etching cycles (for example Ar, Ar / O ₂ , Ar ... or CF ₄ , CO / H ₂ , CF ₄ ... or CO, CO / H ₂ , CO ...).

Although this approach is less advantageous, the two steps F5 and F6 can use two different equipment. The thinning of the copolymer layer may be carried out with a plasma without selectivity with respect to SiARC, PMMA or PS, especially using fluorinated chemistries (eg CF ₄ ), while the removal of PMMA is carried out by other ways than plasma etching, for example wet.

Finally, at step F7 of the figure 1G the patterns 43 in the polystyrene layer 42 are transferred to the underlying substrate 2. The desired patterns 21 (contact holes to form via, sliced to form metal lines, etc.) are then obtained on the surface of the substrate 2. This transfer uses the polystyrene layer 42 and the assembly guide 1 as an etching mask. In other words, only the portions of the substrate 2 located vertically above the holes 43 are etched. Polystyrene 42 and assembly guide 1 are then eliminated so as to keep only the substrate 2 etched at the openings of the guide.

où CD est le diamètre des cylindres et d est la distance entre deux cylindres consécutifs) est égal à 0.01, 0.04 ou 0.2 selon la zone du substrat (soit respectivement 1, 4 ou 16 motifs de guidage répartis sur un même champ de 2 µm x 2 µm).Tests were carried out on a substrate provided with an assembly guide having a height of about 125 nm and a variable opening rate. The guide patterns considered are cylinders of one size. The opening rate, which is equivalent here to a density of patterns $(\mathit{density}=\frac{\mathrm{<figure-callout\; id="4"\; label="\pi "\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">\pi </figure-callout>}}{4}{\left(\frac{\mathit{<figure-callout\; id="2"\; label="CD\; d"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">CD\; </figure-callout>}}{d}\right)}^2,$

where CD is the diameter of the cylinders and d is the distance between two consecutive cylinders) is equal to 0.01, 0.04 or 0.2 depending on the zone of the substrate (ie respectively 1, 4 or 16 guide patterns distributed on the same field of 2 μm x 2 μm).

• conditions de l'essai n°1 : vitesse de rotation d'environ 2000 tours/min, 0.5 % en poids de copolymère dans la solution de dépôt et temps de rotation d'environ 20 s;
• conditions de l'essai n°2 : vitesse de rotation d'environ 1000 tours/min, 1.5 % en poids de copolymère dans la solution de dépôt et temps de rotation d'environ 20 s; et
• conditions de l'essai n°3 : vitesse de rotation d'environ 1000 tours/min, 3 % en poids de copolymère dans la solution de dépôt et temps de rotation d'environ 20 s

For these tests, different conditions of deposition of the copolymer by "spin-coating" have been envisaged:

• conditions of Test No. 1: rotational speed of about 2000 rpm, 0.5% by weight of copolymer in the deposition solution and rotation time of about 20 s;
• conditions of Test No. 2: rotational speed of about 1000 rpm, 1.5% by weight of copolymer in the deposition solution, and rotation time of about 20 seconds; and
• conditions of Test No. 3: rotational speed of about 1000 rpm, 3% by weight of copolymer in the deposition solution and rotation time about 20 s

By way of comparison, the same copolymer was spread with the same deposition conditions on "control" substrates without guide patterns, to determine the equivalent thickness of copolymer on a surface without topography.

Under the conditions of Test No. 1, the equivalent thickness of the copolymer layer (measurable on the control substrate, for example by ellipsometry) is about 12 nm. On the substrate with the assembly guide, it is observed under a scanning electron microscope that certain guide patterns are not completely filled by the copolymer material. The amount of copolymer deposited is therefore insufficient and a flat layer is not reached.

This first test reflects the conditions for obtaining patterns according to the method of the prior art (where there is no planarization step, or thinning step). The thickness values obtained in the guide after assembly differ according to the density of the guiding units: typically 120 nm for a density of patterns of 0.01, 120 nm for a density of 0.04 and 50 nm for a density of 0.2 (AFM measurements) .

Under the conditions of Test No. 2, the equivalent thickness of the copolymer layer is about 50 nm and all the guide patterns are fully covered by the block copolymer. A thinning step, by Ar / O ₂ plasma, succeeds the spin coating and the assembly of the copolymer. It comprises a first etching step of about 20 s to reach the surface of the guide (ie the SiARC layer) and a second step, called over-etching step, of about 5 s to reach the organized portion of the coating layer. copolymer.

After assembly and thinning, the block copolymer layer assembled in the guide measures 110 nm in thickness, where the density of the guiding units is 0.01, 100 nm thick, where the density is 0.04 and 85 nm in thickness. the density is 0.2 (AFM measurements).

Finally, under the conditions of Test No. 3, the equivalent thickness of the copolymer is about 120 nm. All the patterns are also buried by the block copolymer material. The thinning operation following the assembly is identical to that described in Test No. 2, except as regards the duration of the two etching steps: approximately 46 s for the first step and 12 s for the second step step of over-engraving. It can be seen in the AFM images that, after assembly and thinning, the thickness of the block copolymer layer is identical in all the zones of the assembly guide, that is to say whatever the density of the patterns guidance. It is approximately 70 nm.

The figure 2 is a graph of the results of the above tests. It shows the variation of thickness of the block copolymer inside the assembly guide (in percentage of the smallest thickness obtained, ie in the densest units - density = 0.2), according to the density of the guide patterns . The upper curve corresponds to test No. 1, that is to say to conventional spinning deposit conditions (without planarization). The other two curves correspond to the tests n ° 2 and n ° 3, which implement the steps of the Figures 1C and 1E (planarization and thinning).

It can be seen from this graph that the method according to the invention makes it possible to significantly reduce the variation in thickness within a variable density assembly guide. Indeed, by depositing the equivalent of a 50 nm thick layer, the variation in thickness within the guide reaches only 30% (middle curve), compared to 70% in the method of the prior art ( top curve). By depositing the equivalent of a 120 nm thick layer, a uniform thickness i.e. a zero variation) can be achieved. This homogeneity of thickness, obtained thanks to the combination of the planarization and thinning steps, will allow to transfer by engraving both dense contact holes and isolated contact holes, with a minimum of defects (missing contacts or over- etched).

The difference between the 50 nm equivalent thickness curve and the 120 nm equivalent thickness curve is explained by the fact that a better flatness is obtained in the second case (greater thickness, see equation n ° 2). . Nevertheless, as previously discussed, there is already a clear improvement in the thickness final of the assembled copolymer (and thus in the subsequent transfer of the patterns) from 50 nm. Consequently, the deposition conditions may be chosen to deposit a layer of block copolymer having an equivalent thickness greater than 50 nm and preferably greater than 120 nm.

As illustrated on the figure 3A the method for producing patterns according to the prior art causes, after the spin coating, a greater copolymer thickness on the edges of a pattern field 100 than at the center of the same field, because it exists locally. a breakdown of the density of the guide patterns 10 (this density suddenly falls to zero beyond the field 100). However, this difference in thickness usually causes the appearance of undesirable patterns at the edge of the field 100. On the contrary, on the figure 3B , the thickness obtained after the thinning stage of the figure 1E is uniform, both in the center of the pattern field 100 and on its edges. The method according to the invention thus has the advantage of being devoid of this effect this edge.

Naturally, the method for producing patterns according to the invention is not limited to the embodiment which has just been described with reference to the Figures 1A to 1G and many variations can be envisaged.

For example, the spreading of the block copolymer layer in step F3 ( Fig.1C ) can be obtained otherwise than by spin coating, for example by vapor deposition (CVD) or any other technique known to those skilled in the art.

Although this solution is less advantageous, it is possible to thin the block copolymer layer before it is assembled. In this case, the thickness of the block copolymer layer is reduced to a thickness corresponding to the future organized portion of the block copolymer. Preliminary tests have made it possible to determine below what height is the organized portion of copolymer, and thus up to where to thin the copolymer layer.

It is also possible to provide the thinning in two stages, successive or separate chronologically by the step of assembling the block copolymer. In other words, the first of these steps can be performed before assembling the block copolymer and the second step after assembly. For example, an aggressive etching technique will be used during the first step in order to "rough-cut" the copolymer layer, and a less aggressive etching technique for the second step, in order to obtain a thinned copolymer layer with a better state. surface (ie smoother). Moreover, the various modes of thinning described above (CMP, plasma etching, wet etching) can be combined, for example CMP for the first step and plasma etching for the second step.

In addition, other materials than those mentioned above can be used, especially for the layer of block copolymer, for example PS-b-PLA: polystyrene-block-polylactic acid, PS-b-PEO: polystyrene-block-polyoxide ethylene, PS-b-PDMS: polystyrene-block-polydimethylsiloxane, PS-b-PMMA-b-PEO: polystyrene-block-polymethylmethacrylate-block-polyethylene oxide, PS-b-P2VP: polystyrene-block-poly (2-vinylpyridine). Similarly, the hard mask in which the assembly guide is formed can be made of titanium nitride (TiN), silicon nitride (SiN) and / or silicon dioxide (SiO ₂ ), rather than a SiARC stack. / SOC. The number and thickness of the hard mask layers are also likely to vary, depending on the nature of the block copolymer and the etching techniques used. Furthermore, integration of the block copolymer directly into a resin mask can also be envisaged.

Finally, the planarization and thinning steps described above can be used in a hybrid integration combining grapho-epitaxy and chemo-epitaxy, insofar as the substrate has a topography - even very small (greater than 5 nm) - allowing the assembly of the block copolymer.

Claims (17)

Process for producing final patterns (21) on the surface of a substrate (2) by self-assembly of block copolymer, comprising the steps of: forming (F1) an assembly guide (1) on first and second zones (20a, 20b) of the substrate, the assembly guide having, with respect to a reference surface (1a), openings (10) with a opening rate in the first zone (20a) greater than that of the second zone (20b); depositing (F3) a layer of block copolymer (4) on the substrate (2), so that the layer of block copolymer (4) completely fills the assembly guide (1) and forms an extra thickness on the surface reference (1a); - assembling (F4) the block copolymer, resulting in an organized portion (4a) of the block copolymer layer within the openings (10) of the assembly guide (1); - Thinning (F5) uniformly the block copolymer layer (4), until a thickness corresponding to the organized portion (4a) of the block copolymer layer; - removing (F6) one of the phases (41) of the assembled block copolymer, resulting in a plurality of initial patterns (43) extending into the block copolymer layer (42); and - transferring (F7) the initial patterns (43) of the block copolymer layer (42) into the substrate (2) to form said final patterns (21). Process according to claim 1, wherein the deposition (F3) of the block copolymer layer (4) is performed by spin coating. Process according to one of claims 1 and 2, wherein the block copolymer contains at least one styrenic derivative and a methacrylate derivative. Process according to any one of claims 1 to 3, wherein the thinning (F5) of the block copolymer layer (4) is achieved by plasma etching. Process according to Claim 4, in which the thinning (F5) of the block copolymer layer (4) and the elimination (F6) of one of the phases (41) of the assembled block copolymer are carried out in the same frame plasma etching. A process according to claim 5, wherein the thinning (F5) of the block copolymer layer (4) and the removal (F6) of one of the phases (41) of the assembled block copolymer are carried out simultaneously using a single plasma selected from Ar / O ₂ , SF ₆ , C _x F _y , N ₂ / H ₂ , CO / O ₂ , CO / H ₂ , CH ₄ / O ₂ , C _x F _y / O ₂ , CH _x F _y / O ₂ , C _x F _y / H ₂ , CH _x F _y / H ₂ and C _x H _y / H ₂ . A process according to claim 5, wherein the thinning (F5) of the block copolymer layer (4) and the elimination (F6) of one of the phases (41) of the assembled block copolymer are carried out by a succession of stages using plasmas of different types or alternating at least two plasmas. Process according to any one of Claims 1 to 7, in which the thinning (F5) of the copolymer layer (4) is carried out by chemical-mechanical polishing. Method according to any one of claims 1 to 8, wherein the assembly guide (1) is formed of a mask disposed on the substrate (2) and textured by photolithography and / or etching. The method of claim 9, wherein the mask (1) comprises a first carbon layer (11) in contact with the substrate (2) and a second silicon-rich anti-reflective layer (12) disposed on the first carbon layer (11). ), said silicon-rich anti-reflective layer (12) serving as a stop layer during thinning (F5) of the assembled block copolymer layer. A method according to any one of claims 1 to 10, wherein the assembly guide (1) comprises a plurality of openings forming guide patterns (10) distributed between the first and second regions (20a, 20b) of the substrate (2). The method of claim 11, wherein the guide patterns (10) have a height of between 5 nm and 1000 nm. Method according to one of claims 11 and 12, wherein the surface of the guide patterns (10) has a particular affinity for one or more blocks of the block copolymer. Method according to one of Claims 11 and 12, in which the bottom of the guiding units (10) is neutral with respect to the blocks of the block copolymer, and in which the flanks of the guiding units (10) have a particular affinity for one or more blocks of the block copolymer. Method according to one of claims 11 and 12, wherein the surface of the guide patterns (10) is neutral with respect to the blocks of the block copolymer. Method according to one of claims 13 and 14, wherein the particular affinity of the assembly guide (1) for one or more blocks of the block copolymer is obtained by grafting one or more homopolymers. Method according to one of claims 14 and 15, wherein the neutralization of the assembly guide (1) is obtained by grafting a random copolymer.

Images